EXTENDING THE LIFE OF VIRTUAL HERITAGE: REUSE OF TLS POINT CLOUDS IN SYNTHETIC STEREOSCOPIC SPHERICAL IMAGES
J. Garcia Fernandez
a,
, K. Tammi
b
, A. Joutsiniemi
a a
TUT, School of Architecture, Tampere University of Technology, Rakennustalo Building, 33720 Tampere, Finland - jorge.garciafernandez, anssi.joutsiniemitut.fi
b
TAMK,
School
of Construction
and Environmental Engineering,
Tampere University of Applied Sciences,
Kuntokatu 3
, 33520 Tampere, Finland - kalle.tammitamk.fi
Commission II KEY WORDS: Point Cloud, Terrestrial Laser Scanner, Reuse, VR, Stereoscopic, Spherical Images
ABSTRACT:
Recent advances in Terrestrial Laser Scanner TLS, in terms of cost and flexibility, have consolidated this technology as an essential tool for the documentation and digitalization of Cultural Heritage. However, once the TLS data is used, it basically remains stored and
left to waste.How can highly accurate and dense point clouds of the built heritage be processed for its reuse, especially to engage a broader audience? This paper aims to answer this question by a channel that minimizes the need for expert knowledge, while enhancing
the interactivity with the as-built digital data: Virtual Heritage Dissemination through the production of VR content.
Driven by the ProDigiOUs project’s guidelines on data dissemination EU funded, this paper advances in a production path to transform the point cloud into virtual stereoscopic spherical images, taking into account the different visual features that produce depth
perception, and especially those prompting visual fatigue while experiencing the VR content. Finally, we present the results of the Hiedanranta’s scans transformed into stereoscopic spherical animations.
1. INTRODUCTION
During the last decade, laser scanner LIDAR technology has been consolidated as an essential tool for the documentation and
digitalization of Cultural Heritage sites Nothegger Dorninger, 2009. Simultaneously, the cost of the required equipment has
decreased significantly and more versatile software solutions, with advanced processing capabilities, have become available.
The aforementioned factors have ensured the growth of point clouds’ popularity LIDAR native´s data, as a key data structure
in digital Heritage management.
Point clouds are solving a wide range of needs of Cultural Heritage, especially in the areas of conservation and
communication, and loads of digital treasures are therefore being captured and stored. However very few reuse alternatives have
been implemented for generally unstructured point cloud data, while requiring high memory capacities for both management
and storage purposes.
At the same time, from a more social perspective, the effective heritage communication has become a crucial part of the site’s
conservation, as well as an element for the education and cohesion of the society Garcia-Fernandez, 2014. Particularly,
immersive VR content opens a new door not only for widening the access to Heritage-related information, but for a better –
interactive– understanding of complex historical sites Jimenez et al., 2016.
How can highly accurate and dense point clouds of the built heritage be processed for its reuse, especially to engage a
broader audience? This paper aims to answer this question by exploring a channel that minimizes the needs for expert
knowledge, while enhancing the interaction with the as-built digital data: Virtual Heritage Dissemination through the
production of VR content.
2. TERRESTRIAL LASER SCANNER
During the last few years the Laser Scanner Technology has gained a privileged position in the accurate documentation of the
built environment. The substantial decrease in the hardware’s price, the flexibility in terms of dimensions, weight, and scan
speed, together with the increased processing capabilities of desktop computers, has turned terrestrial laser scanners TLS
into a recurrent tool for the built environments documentation Garcia-Fernandez, 2016.
The use of this technology, and its typical data: the point cloud, has already transcended the boundaries of remote-sensing
disciplines. Still, their eminently technical character restricts the use of TLS data to an expert use, especially for the representation,
interpretation and expert analysis of cultural heritage. Multiple documentation projects based on TLS are performed every year
worldwide. Large and rich volumes of data point clouds of cultural heritage assets are stored, at best, once they have met
the specific demands. In a world where the permanence of digital data is so uncertain Carraway, 2011, we are facing the risk to
lose the chances of exploiting the added value of these documents, especially the value for a non-expert public, eager
and in need to take advantage of its cultural heritage Garcia- Fernandez, 2014.
2.1 Reuse of Point Cloud
The concept of point-cloud reuse has been properly addressed in the case of aerial LIDAR Hsu et al., 2011, where national and
international institutions have implemented online-available and free LiDAR Data Sources. However, for terrestrial laser scanner
This contribution has been peer-reviewed. doi:10.5194isprs-archives-XLII-2-W3-317-2017
317
data sites, monuments, and artefacts, only few initiatives have been optimally implemented through open repositories, among
which stands: Large Geometric Models Archive Georgia Institute of Technology; and The Stanford 3D Scanning
Repository Stanford University. The large size of point clouds, hindering both their storage in memory and their transmission
over the network; the diversity of formats; the need for expert knowledge for their use and management, among others causes,
have severely limited the point cloud’s widespread use. Transferring point cloud data into parametric models suitable for
VR content creation often includes complex processes in terms of time and expert effort, such as shape reconstruction features
recognition and parametrization, re-texturization, semantic description, among others Pacheco, 2014; Quattrini, 2016.
Recently, significant improvements in processing capacities, and the opening of modelling and rendering software to point clouds,
allows us to explore the ‘straightforward’ re-use of point cloud in rendered synthetic stereoscopic spherical content, and its
implementation in social communication platforms. 2.2 Our Case ProDigiOUs project
The framework of our study was related to ESF funded project ProDigiOUs http:prodigious.tamk.fiin-english that fosters
the usage of laser scanning in the Architecture, Engineering and Construction sector at Tampere region, Finland. To demonstrate
the potentials of TLS as a built heritage documentation tool, it was agreed with the city of Tampere to implement an
experimental laser scanning of the protected Lielahti mansion at Hiedanranta district.
Driven by the ProDigiOUs project’s guidelines on data dissemination, and looking for answers in TLS’ data reuse that
would meet both the technological and social requirements of easily accessible virtual heritage, we conducted a study aiming to
transform the point cloud into synthetic stereoscopic spherical images, summarized in the following sections.
Figure. 1. Outdoor scan station and point cloud data obtained in in the Lielahti mansion campaign, 2016
The data capture campaign in Lielahti mansion supplied the data to conduct the study cases. It included over 100 scan positions
indoor and outdoor Figure 1; and working in a range of resolution between 1 point every 4-10mm at 10 meters.
3.
SYNTHETIC STEREOSCOPIC SPHERICAL IMAGES
The creation of stereoscopic spherical content taking advantages of the reuse of point clouds, placed us before the opportunity to
recycle and enrich classic stereo photography methods. As in all processes aiming to simulate the depth perception, stereo capture
requires to be mutually consistent, framed in geometric and photometric constraints, in order to optimally reconstruct a 3D-
scene by our brains Devernay Beardsley, 2010. The effects of visual fatigue and visual discomfort when
experiencing VR stereoscopic content has been widely demonstrated by several studies Wann Mon-Williams, 1997;
Atallah et al., 2012; among others. While some standards have been implemented for digital 3D cinema production especially
in terms of format and compatibility with broadcast television standards, there are no regulations for the use and publication of
stereo VR content in public sites. Nowadays, with the development of the digital photography and
the easy access to both software and unverified information, a significant amount of VR content is produced without taking into
consideration any elements that increase visual fatigue such as: disparities, convergence, interpupilar distance, etc. and
consumed by poorly engineered systems. An illustrative case are the multiple websites and free apps announcing the possibility of
doing anaglyph from a single image. Although it is not the purpose of this paper to advance in a
regulationstandards for stereo-VR content, we do propose a serious reflection both for creators and consumers, on the need
to advance in responsible selection and consumption of VR content, and the development of knowledge-based content
creation processes. Within this line, this third section summarizes our progress in the development of a practical guideline in form
of production path for the creation of synthetic stereoscopic spherical images based on point clouds.
The path aims to ensure the visual comfort in 3D-perception while enhancing the stereopsis static perception from fusioning
two binocular images. The path definition takes into account the different visual features related with the depth perception
including the use of monoscopy depth cues, and the aforementioned elements that might cause visual fatigue.
3.1 Defining the path workflow
